基于图像特征的傅里叶梅林变换在图像配准中的应用

Application of Fourier-Merlin Transformation Based on Image Feature in Image Registration

针对相位相关法配准平移旋转的图像具有明显误差的问题,论文中提出了一种基于图像特征进行傅里叶梅林变换,实现对存在平移旋转的图像进行配准的方法。首先对待配准图像和模板图像进行梯度信息提取、二值化、腐蚀、膨胀等运算;然后,将图像的形态学中心平移到视场中心,并对其进行极坐标变换;最后,根据傅里叶梅林变换特性,得到图像的旋转角度。实验结果表明,该方法对存在平移旋转的图像具有很好的配准特性。

Aiming at the problem that phase correlation registration translation and rotation of the image have obvious error,this paper presents a method based on image feature of Fourier-Mellin transform to correct the existence of translation and rotation of the image.Firstly,gradient information extraction、binarization、corrosion、expansion and other operations are carried out to the image to be registered and the template image.Then,the morphological center of the image is translated to the center of the field of view,and the polar coordinate transformation is carried out.Finally,according to the characteristics of Fourier-Mellin transform,the rotation angle of the image is obtained.Experimental results show that the proposed method has good registration of property for the image with translation and rotation.